Commentary: Systematic Review of Safety and Efficacy of Atacicept in Treating Immune-Mediated Disorders

Abstract

A Commentary on Systematic Review of Safety and Efficacy of Atacicept in Treating Immune-Mediated Disorders By Kaegi C, Steiner UC, Wuest B, et al. (2020). Front Immunol. 11:433. doi: 10.3389/fimmu.2020.00433 We read with interest the systematic review article published in Frontiers in Immunology by Kaegi and colleagues, which analyzed information from studies of atacicept across several immune-mediated disorders. Whilst we welcome the effort the authors have made in collating studies of atacicept in different therapy areas, especially the benefit for clinicians and researchers in the field, we have identified several inconsistencies, errors, omissions, and critical flaws in the reporting and interpretation of efficacy and safety. Here, we have highlighted some of the methodological and factual errors in the review (summarized in detail in Table 1) to provide essential balance and context. This response was supported by Merck KGaA, Darmstadt, Germany, who are developing atacicept. Table 1 Details of missing, misleading and incorrect information in the Kaegi et al. systematic review article. The authors identified 10 studies of atacicept in multiple sclerosis (MS), optic neuritis (ON), rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE) suitable for inclusion in their systematic review. The search period was short, from October 2016 to July 2018, and key publications from 2019 were not included. It was claimed that only studies with a minimum number of patients to show a relevant treatment effect were eligible, however, 3 of the 10 studies included were not powered to show clinical treatment effect (1–3) and a further 2 studies did not reach the sample size required for a full evaluation (4, 5). The review was said to be guided by the PRISMA checklist, but there is incomplete or incorrect information provided to meet PRISMA requirements (Table 1). For example, the risk of bias across studies is not assessed and treatment effect measures are not reported in the text. These details are essential for readers to interpret the results correctly. Four SLE studies were included without discussion of the challenging nature of using clinical outcome composite endpoints in this setting, which can result in apparently conflicting results. For example, the primary endpoint in the TULIP-1 trial of anifrolumab using SLE Responder Index-4 was not met (14), whereas the TULIP-2 trial which used the British Isles Lupus Assessment Group-based Combined Lupus Assessment did meet its primary endpoint (15). The atacicept flare prevention trial (APRIL-SLE) provided a novel approach including patients who had recently had a lupus flare that was controlled by a relatively short course of glucocorticoids, but this was not mentioned in the review (12). The analysis of the ADDRESS II primary endpoint focuses on statistical significance, which is misleading as a trend was observed in the 150 mg group (13). SLE is a clinically heterogenous disease and so it is important to identify specific cohorts of patients who may respond to a treatment; the beneficial effect of atacicept in a predefined subpopulation of ADDRESS II patients with high disease activity (HDA, SLEDAI-2K ≥10) (7, 13) was not discussed in the review article. Inaccuracies are also evident in the reporting of efficacy data relating to MS and RA trials for atacicept, as summarized in Table 1. Safety data are reported out of context or with insufficient detail (Table 1). A large safety analysis of atacicept, comprising 17 clinical studies of 1568 subjects and including 761 SLE patients, was not included or discussed (6). The safety profile and number of reported deaths in atacicept studies were found to be comparable with that of other biologic therapies, including belimumab and blisibimod for SLE, but this context was not given in the review article (6, 16–18). Data for atacicept across all studies show that infections and infestations are the most commonly reported treatment-emergent adverse event (45.6%) (6). This is not unexpected since atacicept reduces immunoglobulin levels and B and plasma cell numbers, and is consistent with other biologic agents used to treat autoimmune diseases (6, 19). Overall, atacicept is associated with increased infection rates compared with placebo, however, serious and severe infections are not higher with atacicept in patients with SLE, RA or ON (6). The authors correctly report that two infection-related deaths occurred in the 150 mg arm of the APRIL-SLE trial, but this led to discontinuation of the 150 mg arm only, not the whole trial as stated (12). Unfortunately, many trials in SLE record a small number of deaths. The APRIL-LN study in SLE was stopped with six patients enrolled due to a decline in serum IgG and the occurrence of serious infections (4). On further analysis the decline in IgG levels was linked to the mycophenolate prescribed prior to the addition of atacicept. A risk mitigation strategy was implemented for subsequent studies; in the Phase II ADDRESS II study of over 300 SLE patients, infection rates were lower and no deaths associated with atacicept were reported (13). Therefore, with the implementation of effective mitigation measures to reduce the risk of infection, the benefit of atacicept for SLE patients with HDA may outweigh the risks (6). It is imperative that this is highlighted in the review article. Kaegi et al. conclude that atacicept failed to show superior effect on disease activity in comparison to placebo in MS, ON, RA and SLE without inclusion of all relevant data, especially in the case of SLE, or full acknowledgement of the limitations of the review. In fact, in all studies, atacicept did show an effect on disease activity (as indicated by a reduction in biomarkers) but this was not always translated to measurable clinical efficacy over...